Hyperfine, Inc. (HYPR) Earnings Call Transcript & Summary

The Webinar will now begin.

Good afternoon, and welcome. I'm Eddie Knopp. I'm the Vice President of Medical Affairs here at Hyperfine. And I'd like to thank everyone for joining Hyperfine webinar symposium on the value of Portable ultra-low-field brain MR imaging. This afternoon we have 3 fantastic speakers, that are basically here to show you the value of the Hyperfine Swoop System. First up, we're going to start with Dr. Jamal Derakhshan, who comes from the Jefferson healthcare system. The Abington Hospital where he's going to relate to you the clinical utility of the system in a basic practice. The Jefferson Abington site is the largest commercial clinical site that we have, scanned in excess of 300 patients. Next up, Dr. Taylor Kimberly, who is an Associate Professor at Harvard, and at the Massachusetts General Brigham Hospital, where he runs Neurocritical Care and the stroke service. And Taylor will talk to us about the utility of the system in the setting of Stroke. And finally, we'll end with Dr. Cyrus Raji, who comes to us from the Washington University, St. Loius, Mallinckrodt Institute of Radiology, Barnes Jewish healthcare system, and Cyrus is going to speak to us about the utility of the Swoop scanner in the setting of dementia and Alzheimer's disease. With that, I'll turn it over to Dr. Derakhshan, who will speak to us about his clinical experience. Jamal.

All right. Thanks very much for inviting me, Dr. Knopp and for the kind Introduction. My name is Jamal Derakhshan, I'm a Neuroradiologist. And this work was done at Jefferson Abington Hospital as part of a private practice that I'm in, called the Radiology Group of Abington, we'd like to share that experience with you. And I'll thank all my colleagues at the end of this talk, who also contributed. Next slide please. So what I'm showing you here is some of the differences between a standard MRI and the portable MRI that's been developed by Hyperfine because it really is a huge leap in technology, which I think is very important. So on the left-hand side of this slide, you have a picture of myself standing in the control room of a 3 Tesla scanner. And you can see that there is a very important site restrictions in terms of building a room that's appropriately shielded for a standard MRI and the equipment is massive. The scanner weighs about 10,000 kilograms. It requires superconducting magnet, which requires a bath of liquid helium that has to be replenished in specific power sources, you have to monitor the room for depletion of oxygen, which is the arrow that you see there showing the oxygen level in the room. So really, this has been a huge leap in MR technology being able to miniaturize all of this equipment down into a portable unit, which you see on the right-hand side, which can be wheeled around the hospital and taken to the patient. So this is really a paradigm shift that opens up a lot of avenues for scanning patients both, as I'm going to show you in a tertiary care hospital, but also in underserved areas, remote areas and other places in the world. So this is a very important technology. Just to show you some of the differences, I'll go back there real quick. So we have 20x lower magnetic field strength. So it's only 4% that they're able to use to generate pretty good images. We use much less RF energy. So issues with heating that would be present in a standard scanner are really not present because we're only using [ 0.02% ] energy. The noise is much lower, so we don't have to worry about acoustic noise, damage to patients. And we bring the patient -- we bring the scanner to the patients instead of having to move the patient, which is very critical in some applications, where patients are being monitored in an intensive care unit. And this leads to improved patient safety. Next slide, please. So I'm going to go through our experience using this 0.064 Tesla brain-only Hyperfine Swoop scanner. We've, as Eddie mentioned, we've scanned over 300 patients at this time. At our institution, we trained X-ray technologists to be able to use this, which is the difference. X-ray technologists are more available than more specially trained MR technologists. We basically screened the patients like we would for routine MRI in any questions that if they came up with them by MR a medical director, like myself, ICUs patients were scanned in the ICU and the floor-in ER patients were actually brought to the scanner and the CT holding [ pages ] so there were additional people around monitoring the patient. And these studies were read by 4 fellowship trained neuroradiologists for the majority. Next slide, please. So this is an example of a patient setup in the ICU room, where the Swoop MRI has been brought into the ICU. The patient's bed has been rotated 90 degrees, and the patient has been moved into the scanner. And the important thing here is that a lot of the ICU equipment, you can see around the patient in terms of infusion pumps, those can be kept in place and they don't have to be moved. And they're outside of the so-called 5 gauss line, which is denoted by the orange ring that extends. And basically, anything outside that, you can have any of the support equipment that would be needed, available. And certain things that need to go in like EKG monitors and things, are that it's much safer to do that with the improved safety characteristics of this scanner. Our routine protocol was 37 minutes. It involved what we call a localizer. And then these are specific pulse sequences and Axial T2, Axial and Coronal FLAIR imaging, Sagittal T1 and Axial diffusion with what we call an ADC map. So these are 5 flavors of sequences that we run, and we're able to get a lot of information using these. Next slide, please. So when we reviewed our first 100 patients, we had a nice balance of men and women, 52 men and 48 women. These were predominantly older patients, 65-year-olds, plus or minus 20 years, but we actually scan patients from several days old, all the way up to 97 years old, and we found it to be useful in all of those categories. We did scan patients using implants, I have a background in MR physics. So I was able to contribute to knowing what we were able to do, and thankfully, all of this was done safely. There was a patient with a non MR conditional pacemaker that had a brain abscess and we were able to scan them 3 times. We had a non-MR-conditional spinal stimulator due to an abandoned lead that we were able to scan safely. And we also scan devices that have been approved for 1.5 Tesla on the low-field scanner safely as well and those involve 2 pacemakers and 1 spinal stimulator. We scanned a bunch of passive -- patients with passive implants. These days, it's very common for patients to have either passive or active implants. And you can see the list there of the various objects which were in these patients, and were able to be scanned safely. And I'll highlight 2 patients. One of them was a man who had, had 3 metallic devices placed in their head as a child, after a motor vehicle collision, we weren't sure about the safety of those, and we were able to scan them safely on this lower field scanner. There's also a patient with shrapnel in their legs and some other patients that were able to be scanned safely. We did administer contrast to 2 patients because they were not able to get a high field scan. So that was a patient with a brain abscess and also with suspected meningitis. Next slide, please. This is the example of the patient with the brain abscess in their right thalamus. On the left-hand side, you can see the diffusion-weighted image, and we can see the characteristic restricted diffusion in the residual abscess. On the middle image, you can see a rim of enhancement. And these are the first images that I'm aware of on the low-field scanner giving contrast and showing that we're able to detect contrast with a standard dose of IV gadolinium using the scanner. Based on that, the -- they may pursue post contrast imaging labeling on this as well, that's currently FDA-approved for non contrast brain MRIs. So this was done off-label. And in the right most image, you can see the healing abscess, where the cavity has involuted, and basically using this, they were able to prescribe the appropriate antibiotics and also know when to stop the antibiotics when that abscess was completely treated. Next slide, please. In this patient who was found after strangulation, we can see restricted diffusion in the right thalamus with the arrows. And then more generally, we can see increased T2 signal throughout the gray matter in the brain. And this is indicative of significant brain injury. In fact, this patient later went on to have a catheter angiogram, was found to be brain-dead with lack of blood flow to the head. So this was able to be used to prognosticate for the patient in the ICU without having to move them to the scanner. We can go ahead and go to the next slide. This was an example of how the Swoop is very useful in a tertiary care setting. This is a highly critical patient that had unfortunately a embolus in their basilar artery, which is the main artery supplying the brainstem. You can see it on the left catheter angiogram, there is a filling defect. On the second image, you can see an unfortunate complication where the guidewire that's been used to go up and try to remove the clot has caused the vessel to rupture and that's iodine outside. And so the catheter was immediately removed at that time. The patient was taken to the ICU, a very critically ill patient. They actually requested the Swoop to be brought, so they could prognosticate and the Swoop showed an infarct unfortunately, in the right cerebellar hemisphere in the third image. And then on the fourth image, we can see blood in the subarachnoid space as well as in the ventricles. So basically, this is able to help them understand the extent of injury and determine next steps for the patient without having to move the patient from the location that they are being highly monitored. Next slide, please. This is an example in an infant. We can see there is some layering hemorrhage in the left occipital horn, shown on the -- on the images here, importantly, we can get both axial and coronal images, the multiplanar feature of MRI on the scanner. And we did detect this hemorrhage, which, in retrospect, we can see the other images in ultrasound that's used very frequently in young children. We can in retrospect, see a little bit of hemorrhage there in the occipital horn that had been missed on the ultrasound was picked up on this MRI. Now we can go to the next slide. This was another infant. This was a 3-month old that had known hydrocephalus that have been monitored with ultrasound -- serial ultrasound. This is a result of what we call germinal matrix hemorrhage. It's hemorrhage in the portion of the brain that's common in patients that are premature. We can see the blood in the lateral ventricle and the fact that the ventricles are enlarged, so-called hydrocephalus, which they knew about, but we gained additional information. For example, in the left cerebellum, there's a wedge shaped area of tissue missing, and that's called encephalomalacia. That's a result of a prior injury. And this was able to detect that and it had been missed on ultrasounds. I'm showing you the ultrasound on the right side. And what you can see is that the ultrasound beam does not penetrate into the -- what we call the posterior fossa, were around the brain stem. And that's just a limitation of how far the ultrasound can see through the Fontanelle that's open. And this MRI allows us to access more information on that patient, which will help the patient, the team know how to treat that patient more appropriately and know the expectations for their ultimate injury and recovery. You can go to the next slide. This is a patient that was clinically seizing. There's always a question of, is it seizure real or is it what we call a pseudoseizure. In this case, the patient was, I believe, in status epilepticus, which means that they need to be critically monitored, the MRI was brought to the patient in the ICU. And we can see the restricted diffusion in the left hippocampus shown by the white arrows. The left image is the FLAIR image, the middle image is the diffusion weighted image and the right image is what we call the diffusion coefficient, the ADC, the apparent diffusion coefficient. The reason I included this slide is this is a more subtle diagnosis that we make with high-field MRIs, and this just shows you that the Swoop image quality has been improved to the point where we can start to see these things that we consider more subtle findings on MRI using this portable MRI device. So we can go to the next slide, please. This was a critically ill patient. It was a 20-year-old motorcyclist. He had been thrown 50 to 60 feet from the motorcycle, Unfortunately, I believe he was run into by a garbage truck. We can see there's a lot of scalp soft tissue swelling, we can see a little bit of layering hemorrhage in the right occipital horn with the dashed arrow there. But importantly, the image on the left there is information that we would have been able to understand with a head CT scan. But the image on the right is something that you can only see with an MRI. And this shows bright signal, which is restricted diffusion and what we call the splenium, or the posterior aspect of the corpus callosum, which is white matter tracts that connect the 2 hemispheres. And we can see that the water molecule is not diffusing appropriately, and that's because of what we call diffuse axonal injury. And that's a very important finding. It explains why this patient was partially comatose and it has implications for their recovery, how they need to be managed. And whether they're going to be coming out of their injury very quickly or it may take a little bit longer. And that information was available using the Swoop. Thank you. Got to the -- yes. And this is another example of the utility of the Swoop. So we were able to use the physical characteristics of the scanner. It's very useful to -- and this is an example of why even in the United States, in a place that has many high-field scanners. It's useful to have one of these scanners in a hospital. This is an example of a patient that had dental work, which is very common. That dental work has metal and that can distort the magnetic field. And what we can see on the FLAIR image on the left is that we've lost the suppression of FLAIR signal in the sulci, where the places between the gyri [ entirely. ] And that's due to an inhomogeneous magnetic field. Somebody may be concerned that meningitis could look that way or hemorrhage could look that way. When we look at the diffusion weighted image, the image is very blurry and these are the images at high fields. So this is the kind of clinical gold standard. So when I read this, I called this a nondiagnostic study. The patient actually got transferred from our secondary hospital, Landsdale Hospital to Abington Hospital, and they actually got an image with the Swoop. And the important thing is that, that susceptibility artifact scales with the magnetic field. So because the magnetic field strength is 20x less strong, that artifact is not as significant, and we can see that on the FLAIR image, we have the normal suppression. And on the diffusion, we can access information about diffusion and know if there's been a stroke or not, and that patient is able to get to their ultimate diagnosis using this device, which was unable to be obtained with the standard high-field scan. So what we find is that it's very useful to have one of these devices around to scan patients that may have either implants or other things that prevent a standard high field scan. We can go to the next slide, please. This is an example of a patient that had dementia. One of our later speakers is going to give a much more complete talk about dementia. This -- the only thing to note is that we picked up a subtle skull-based mass and the right cerebellopontine angle, that in retrospect, you can see on the head CT scan, but it kind of -- it blends in with the brain stem. So it hadn't been called on that head CT. And here, we could see it very clearly with this brain MRI. So again, added value of this Swoop MRI compared to a CT scan. We can go ahead and go to the next slide, please. So overall, just kind of summarizing our results. Our overall scan time was about 40 minutes. 54 of the patients had acute findings. Only 24% of them required an high-field MRI the same admissions. Some of those were obtained before the Swoop only four other patients had contrast. The scan time was longer, but we did see that there's very avid enhancement in the nasal turbinates and also in that abscess. And 89% of them were considered to be diagnostic, 5% had small undetected infarcts that we found [ in Highfield, ] and some of them had severe diffusion artifacts, only 2 of them. and only one of them was the diffusion considered and conclusive. 3 of them had motion. We can go to the next slide. So in conclusion, we think that portable MRI is a new clinical use for imaging modality. When I ask our clinicians, what they felt? They felt that it added convenience to be able to scan the patient right in their ICU room and not have to move them. It allows them to get a faster scan. And we can use it to scan patients who can't be moved and patients with the susceptibility artifacts or non MR compatible devices. We can readily see intracranial hemorrhage, hydrocephalus and significant infarcts, and we can see contrast enhancement. And we do demonstrate that we can scan patients safely with active and passive implants at 0.64 Tesla. So I'd just like to thank my colleagues there at Abington Hospital, our technologists who did it, the other neuroradiologists who are reading and our Hyperfine colleagues, as well as Dr. Shelat from USC, who's an expert on MR safety that I collaborated with. Thank you very much.

Thank you so much Jamal. What a fascinating talk. One quick reminder to the participants, please put your questions in the question-and-answer box so that we can have them addressed at the conclusion. Now I'd like to turn it over to Dr. Taylor Kimberly, and he's going to talk to us about stroke, Taylor.

Thanks, Eddie, for the introduction and the opportunity to speak about our experience with the Hyperfine Swoop. Just as background, I'm a neurologist and neurointensivist. And about 5 years ago, when I started working with this device, began in the setting of the ICU, where bringing imaging to the bed side is so critically important. And even back then, it became clear that there were many other a broader use case scenarios. But today, I'll talk about one of those, and that's positioning the Swoop in acute stroke workflows. So if we go to the next slide, as Dr. Derakhshan had mentioned in the beginning of his presentation, the development of ultra low-field MR really is a paradigm shift. As a clinical provider at the bed side, the ability to bring a diagnostic quality imaging and neuro imaging device to the patient opens up a huge variety of really intriguing patient-centered applications. As was mentioned in the first half, the usual standard of care is to bring a patient to a highly access controlled environment where there's exquisite image quality. But by miniaturizing many elements of the scanner and putting it on motorized wheels, that allows a scanner to be brought to the patient at point of care, both in inpatient and outpatient settings. And in ways that I think will transform how we think about medical care going forward. It can also be brought out into the community, either to community health centers given a safety profile and ease of use, as well as in the pre-hospital setting and then other scenarios, including in low resource rural or low and middle income countries. If you go to the next slide, the scenarios that we focus on have been primarily on the inpatient and outpatient realm. And today, I'll just focus on some of our work in acute stroke. A lot of the work that we've done has been a collaboration with several other centers, including University of Buffalo, Ohio State and University of [ Glasgow, ] as well as MGH where my team is located. And -- so I'll focus on some of our work, primarily in the ED setting, using Swoop. Next slide. So the next slide. So -- just wanted to start out with a couple of case examples just to show the value and sensitivity of the scanner in its current state. This is using the latest generation hardware and software. So it's a 74-year-old woman and she presented with pretty typical stroke symptoms, right-side weakness, aphasia, where she was unable to generate words. And accordingly, she had an NIH stroke scale score at 16, which is consistent with the severe stroke. We have the unusual scenario even for academic medical centers for having a scanner in our -- a high-field conventional MRI scanner in our ED and that really allows us to compare side-by-side the performance of the scanner -- of the Swoop to the conventional MRI scanner. The top 3 images are from that conventional 3 Tesla scanner, the DWI or stroke sequences on the left showing that bright white region and an accordingly dark ADC image in a slightly bright FLAIR image. And on the Swoop scanner, it's in fact, quite similar, same location, same sensitivity. Next slide. And it's not just limited to these severe strokes, but we've found good use cases where the stroke is mild and the lesion is smaller. And in this case, a similarly aged man presented with just some slurred speech and a very mild stroke with a NIH stroke scale of 1. And you can see on the top 3 images that this corresponded to a very small acute stroke that was about the size of a pencil eraser. And indeed, the Swoop was able to detect that on the DWI acute stroke sequence as appropriately dark on the ADC and detectable on the FLAIR. So it has sensitivity across the range of stroke severity. So when thinking about how to position the Swoop scanner within the landscape of stroke care, there are a number of potential points at which it can be inserted into these workflows. It can be used for initial triage, for distinguishing hemorrhage from ischemia, and there are a couple of scenarios where that could be useful in the delayed window evaluation. So there are a number of patients who present outside of the traditional 3 to 4.5 hour time window, which is a typical threshold for IV thrombolysis therapy and where MR imaging can be used to characterize the age of the stroke. Similarly, for those who wake up with a stroke where there's an uncertain type of onset, the imaging characteristics, which I'll go into in a little more detail in a few slides can be used as a surrogate for a time of onset, that can be very helpful for treatment decision-making around IV thrombolysis. And then the Swoop scanner can be used for posttreatment monitoring. And there are a number of performance characteristics that this Swoop has that make each of these indications feasible. Next slide. So one of the things to recognize is that acute stroke algorithms have also increased in complexity, especially over the last 5 years, as endovascular thrombectomy has been demonstrated to be quite effective in conjunction with IV thrombolysis. But in each of these scenarios, and as a central tenant of acute stroke therapy is that time is of the essence for our diagnosis, treatment decisions and then monitoring thereafter. And one of the core features is not only keeping track of the clock and the time that's passed since the stroke onset, but really using a neuroimaging characteristics to guide these treatment decisions. Next slide. So one of the use cases that I'll highlight here, which I think is a very intriguing application is the concept of a wake-up stroke. So there's about 25% of stroke patients will go to bed totally fine. And upon awakening, they will have symptoms. Now evidence suggests that a majority of those patients have had their stroke onset in the latter part of that -- the sleep period, and in many cases, actually, it's the event that prompts awakening. But because of the risks associated with TPA, particularly outside of the 3- to 4.5-hour time window, those patients typically don't qualify for treatment. Over in Europe, in France, where acute MR imaging at 1.5 or 3 T is used as part of standard of care -- they ran a study using characteristics on the MRI as a surrogate, where it served as a tissue clock as compared to the traditional time clock. And the concept behind this is that if the stroke was visible on the DWI or stroke sequence, but not detected on the T2 FLAIR sequence that mismatch was a marker for a stroke that was early in the time window and could benefit from treatment. Next slide. And in fact, this has been proven in a randomized controlled trial conducted in Germany and France showing that using MR to guide IV thrombolysis when the stroke onset time is unknown. It improves outcomes. And this graph shows a greater -- a percentage of the lighter boxes, which correspond to better outcome among those patients randomized to alteplase as compared to those randomized to placebo where there are a greater proportion of darker boxes corresponding to poor outcome. So the concept of using MRI to guide thrombolysis was established in this study. But the challenge is, is that it was not generalizable that the accessibility to conventional MR is just not there for a vast majority of not only the U.S. but the world as a whole, with the exception of these very few sites in Germany and France. So this really represents a golden opportunity for the Swoop to fit this well-recognized need and help define treatment decision-making. Next slide. So to demonstrate evidence that the Swoop scanner can serve this purpose dating back to 2020, when we were scanning primarily in the ICU, we demonstrated the proof of principle that the Swoop scanner was sensitive to the detection of acute ischemic stroke. Next slide. And that in a series of -- in a case series of acute stroke patients in which we scanned with the Swoop scanner, we were able to demonstrate that the FLAIR DWI mismatch was an excellent surrogate. In this case, the 2 images on the left show DWI stroke sequence on top, and that FLAIR sequence on the bottom. And in a patient in whom the stroke onset is known. So it's within a 3-hour time window when thrombolysis would be -- a patient would be eligible and that stroke lesion in the left hemisphere is FLAIR dark. And in a separate example for a patient who presented 12.5 and then another one who presented 21 hours later, that FLAIR -- that lesion is not only FLAIR bright, but also DWI bright. So this concept is recapitulated with the Swoop scanner has a tissue clock. Next slide. Just back one slide. And this -- we systematically studied a cohort of approximately 50 patients, with acute stroke. And this is in a multicenter study with our collaborators. And this shows in principle that that threshold of increasing FLAIR signal intensity is an excellent marker for their tissue clock, using the Swoop and this actually exactly recirculates some of the information that had been established previously with the conventional MR. Next slide. And so where we're continuing to go and where Hyperfine continues to go is to improve the image quality even further. And I'll say, having the privilege of work with the Hyperfine team over the last 5 or so years, it has been remarkable how quickly the image quality has increased, and it has not stopped. And this principle applies for acute stroke detection as well, where now the technology is able to acquire a DWI stroke sequence in 3 directions, which has the added benefit of increasing the lesion conspicuity which you can see in these 2 right-hand images, where the overall intensity of the brain in more homogeneous and can be easily be distinguished by eye from the stroke as well as on the ADC. I show another example on the next slide. In this case, it's on the -- it's the left-hand -- I should say it's on the left-hand side of the picture, but it's the right hand -- the right side of the hemisphere, right side of the brain, that bright lesion, while detectable in each direction of the X, Y and Z plains, once that trace image has compiled the lesion constituting increases as well as the darkness on the ADC. This is a characteristic pattern of an acute stroke. Next slide. So in summary, Swoop has a potential role in acute stroke evaluation at several steps. I highlighted one example, which is the wake-up stroke example because it is an unmet need where there is an FDA-approved treatment that we know works. But the conventional MR is not routinely available for a vast majority of patients. So this really identifies a use case for Swoop, where it can be used as a tissue clock and a surrogate for the stroke time of onset. And as I mentioned earlier, there are a variety of other really intriguing indications, including in mobile stroke units in the pre-hospital setting, together with colleagues in our emergency department, we've talked about its value as a triage tool for low-risk TIA and then also in the setting of interhospital triage and transfers in the setting of hospital systems. Next slide. So that's it. So thank you very much.

Thank you very much, Taylor. So once again, a reminder, any questions, please put them in the question-and-answer section of Zoom, and we're going to move on to our last speaker. Dr. Cyrus Raji, who's going to talk about the need for Hyperfine in Alzheimer's disease neuroimaging. Cyrus, the floor is yours.

Thank you, Dr. Knopp, for the kind introduction and to the other speakers for giving us important perspectives about the need for Hyperfine Swoop in community and academic neurology settings. I'll be speaking to you today about why we need Hyperfine for Alzheimer's disease neuroimaging? So let's go to the next slide. And what I'm showing you on this slide is we have a variety of different techniques for imaging Alzheimer's disease. And -- when we look at the acute ICU setting, for instance, there are about 5 million patients or so who go to the ICU every year, and quite a few of them will want the MRI imaging. For Alzheimer's disease, which currently affects close to 7 million individuals in the United States and projected to increase to 14 million by 2050, every one of these individuals will need multiple MRI scans. And we see that we have different weightings or assessments to look at MRI pathology and anatomy. We can look at advanced imaging techniques like diffusion tensor imaging, and this alphabet soup is simply showing you all the different ways in which we can characterize the brain's anatomy, physiology as well as using PET scans to look at the molecular proteins and abnormalities metabolically of Alzheimer's disease. And so we have a lot of ways to characterize Alzheimer's disease with imaging that have been validated in the literature. But as I'm going to show you on the next slide, we really have to start thinking about the scaling issues and challenges in terms of the number of MRI scans we would have to do to really serve this population. And so -- what I'm showing you here is an estimate of the number of total MRI scans done in the United States in 2016 and 2017. It is basically between 36 and 39 million total MRI scans. The majority of which are done in nonhospital locations or just the kind of locations that will be served by a mobile deployable unit like a Hyperfine Swoop. And as we see on the next slide for perspective purposes, we know that PET scans, which are often used to image oncologic cases are very reduced in their scale when it comes to the number of scans we can do. And so what you see here from 2003 to 2020 is the number of PET scans does increase substantially over time. But as of 2020, we can only do a little north of 2 million PET scans in the United States. So for scalability purposes, we really are going to be served in the Alzheimer's community by focusing on MRI as opposed to PET. But given that we have 7 million individuals with dementia and Alzheimer's in this country and that is projected to increase to 14 million by 2050, then if we need to do at least 3 MRI scans on each individual in 2050, we're going to need 42 million additional MRI scans that we currently aren't doing. And so having platforms that can increase our reach, affordability and deployability of MRI are going to be super important for the success in serving this population. And so as we move to the next slide, we can accomplish this goal with Hyperfine because as we see here, the Hyperfine Swoop system literally allows you to wheel the MRI scan to wherever the patient needs to get the imaging and whether it's the intensive care unit or as I'll talk about an infusion center for the use of Alzheimer's treatment for anti-amyloid agents, having the sort of flexibility as well as the very low field strict, which, of course, allows us to scan individuals with all kinds of medical hardware that would not be as easy to do with high field MRI will be incredibly useful. And as we move forward to the next slide, it's important to understand that the need for imaging in Alzheimer's disease with MRI is going up because we now have new drugs that allow us to track the clearance of the causative protein of Alzheimer's disease. And so what you're looking here are 2 rows of PET scans. One row is a PET scan of somebody with Alzheimer's disease. And all the red you're seeing is the amyloid proteins. These are the proteins that deposit in the brain, decades before the symptoms show up. And as you can see, once they deposit, they tend to hang around and cause a whole bunch of downstream neurotoxicity, loss of neurons, loss of synapsis, reduced connectivity in the brain. And this, of course, leads to the cognitive symptoms and memory loss of Alzheimer's disease that are progressive and debilitating. But with the new anti-amyloid immunotherapies that have been developed that I'll talk about shortly, we can actually clear amyloid from the brain. So what you're seeing here in a treated participant is positive PET scan with amyloid, all the red we're seeing showing the abnormally high Alzheimer's proteins. And then we can see that red gets reduced with treatment. And this is really imaging the treatment responses. But as we'll note and going on to the next slide, that one of the consequences of treatment is that we can often get amyloid-related imaging abnormalities. These are findings of edema and hemorrhage that can occur and are best track with MRI scans in relationship to treatment. So these treatments do slow memory loss. They don't completely cure the disease, but they slow memory loss by actually acting on the pathology in the way the previous drugs haven't done but they do lead to these complications called ARIA that we now have to track. And we need at least in a given year with treatment to 3 MRI scans in order to evaluate potential treatment responses. And so as we go to the next slide, we can get a better understanding of how these ARIA abnormalities are characterized. And again, the 2 main flavors are edema and hemorrhage. With edema, we need to look at actual abnormal swelling in the brains tissue, the fusions in the brain sulci, which are the spaces or gaps between the gyri of the brain. And as you can see, depending on how big that FLAIR abnormality is or that abnormal brightness on the scan is we can characterize the ARIA-E as mild, moderate or severe. And so too, similarly for hemorrhages, based on the number of hemorrhages we can track on a scan, we're going to note that you can have a mild, moderate or severe hemorrhage. And even though hemorrhage sounds really scary, it's actually the edema of ARIA that is more clinically important to track and more clinically consequential and symptomatic in the ARIA-H. And we can get a better understanding of that. Moving on to the next slide, where we can see the 3 main drugs that are currently either FDA-cleared or to be imminently FDA cleared and they are aducanumab on the top left of the screen, lecanemab, which is the right, which is currently FDA-cleared. And then, of course, donanemab, which is the more recent drug under consideration, the FDA advisory panel gave unanimous recommendation for the clearance, but it's currently pending. And these 3 drugs have the same action. They're all infusion, so you need to actually inject somebody at an infusion center. These aren't [ in tell form ] but they do have a couple of things in common. In addition to their mechanism of action, they all lead to a certain percentage of ARIA-E and ARIA-H. But as we look at the highlighted portions from these papers, for example, with lecanemab, which is the most commonly used from of the drug right now. We see that there is a difference in the amount of individuals we get ARIA-E with symptoms versus ARIA-H. And so for instance, in this trial, what they noted is that ARIA-E is certain certainly common. ARIA-H is a little more common, but ARIA-E has a greater amount of symptomatic individuals. And so in the lecanemab trial, 12.6% individuals got ARIA-E versus 17.3% for ARIA-H in the treatment group compared to placebo. And in the ARIA-E group, we can see that there were 2.8% of participants with symptoms versus 0.7% symptomatic isolated ARIA-H. And so that's an order of magnitude difference in the symptoms. And so to really track the more clinically important variety of ARIA, it's really the edema that we have to focus on, especially when we have somebody on treatment who comes in with symptoms into ER settings, which does happen. And of course, we can see that there are similar data with donanemab and aducanumab. Interestingly enough with aducanumab, it does not reduce the actual symptoms of Alzheimer's disease, so it's not as commonly used as lecanemab, which does lead to cognitive slowing, reduction. And then donanemab, there's also blood, the cognitive deterioration of Alzheimer's. But as I said, it's not FDA cleared yet. However, the panel has approved this clearance. So as we move on to the next slide, knowing that we actually have to track ARIA-E and ARIA-H , but also understanding that ARIA-E is the more clinically important of the 2. I wanted to show you some examples from the literature of what ARIA-E looks like on FLAIR images and what ARIA-H looks like. And what we can see on the left of the screen is that ARIA-E manifests as these abnormally blocky, bright areas on 2D axial FLAIR imaging. And so -- this is high field. This is 1.5 T, but it's not the highest field we could get clinically. It's not 3 Tesla. And we can see these are 2D sequences, so they're not advanced 3D sequences. And we can see clearly that this abnormally high edema is noted with the bright signal on the FLAIR scan in the -- using radiological right, the right hemisphere of the image in the frontal lobe and higher up. And we can see on the right-hand side of the screen, gradient echo sequences showing ARIA-H, which were to be small, often tiny hemorrhages in these linear areas of siderosis and the sulci. So again, the most common type of ARIA , but not as clinically consequential as ARIA-E as we know on the left side of our screen. So when going to the next slide, we can also understand that the importance of ARIA-E compared to ARIA-H is that ARIA-E often goes away. It appears with treatment. It goes away when treatment is stopped. And then when post dosing is followed up, we can actually see that it goes away. When hemorrhages occur then they're no longer going away. They're pretty persistent findings. And why is that important? Because if we stop the drug and the ARIA-E goes away, if the edema goes away, that can be an indication to restart the drug. So this is another reason why ARIA-E is more important to track compared to ARIA-H because not only is ARIA-E more symptomatic, but if it goes away ,which it often does, then we can actually potentially restart the treatment in the patient, especially if the symptoms were on the milder side. And so being able to track this on longitudinal scans as we note here from this paper in the AJNR that we published is actually very important. And so as we go to the next slide, we can see that we have a potential paradigm in which given the portability of Hyperfine Swoop, it can be paired with these infusion centers. And of course, it's estimated that to meet the need for all of the Alzheimer's treatment, given the 5 million or so individuals or 7 million or so who are currently having dementia and would need treatment. You're going to need thousands of infusion centers, thousands of clinics doing this, and you could pair Hyperfine with multiple of these centers and thus have the ability to track at the very least, the more clinically important aspect of ARIA. And that's not to say that you wouldn't want to use high-field MRI, I don't want to suggest that you would necessarily use Hyperfine Swoop to replace high-field MRI. I would think of it more as using the Hyperfine Swoop to replace CAT scans. Which would characterize the edema as sensitively and of course, involve the use of radiation. And so if you can actually expand your MRI footprint to track ARIA-E and pair it up in this kind of infusion model paradigm that will be really important for actually expanding the use of this therapy. And then if you wanted to get the higher field scan to track hemorrhages, that is certainly an option as well. But when somebody comes in with symptoms, getting that answer more rapidly, I think, is going to be quite amenable to this approach with Hyperfine. And so, as we go to the next slide, we can start to see how we're doing a lot of work in this area. So Tammie Benzinger, here at Washington University in St. Louis, who's in Mallinckrodt, Barnes and a real pioneering leader in Alzheimer's disease imaging, including understanding ARIA as well as portable MRI, has been partnering with Dr. Gomez Isla at the Massachusetts ADRC of Harvard Medical School to really study if we can use the ability of Swoop to detect ARIA-E to better track and help these patients, especially as they need multiple scans over time to evaluate the potential safety risks in ARIA findings. And of course, in this study, they'll get both high field and super [indiscernible] if there is a covalent of tracking the ARIA-E in the study. And so that's a very exciting state that it's going on right now. As we move to the next slide, we can also see examples already of a Hyperfine scan on your left, showing severe FLAIR white matter hyper intensities. And of course, if you had edema from ARIA-E, you could have the ability to track similar findings on the scan. And of course, the same patient got a 3 Tesla MRI scan the same week, as you can see, we can better track the FLAIR hyper intensities. By the way, tracking these FLAIR hyper-intensities in this case, the result of chronic small vessel ischemic disease is important for its own reason because this can be a contraindication to getting on treatment. And so knowing this at baseline has a lot of clinical implications for who gets treated. And of course, small vessel ischemic disease is an important metric of brain health overall and because the more of these white matter hyper-intensities you have from this cause the less healthier of a brain you actually have. And so as we go to the next slide, we can appreciate that on some work that I did a few years ago, we actually correlated the amount of white matter hyper intensities. These bright spots that you just saw on the previous screen. And what we found was, not only does more white matter hyper-intensities predict more atrophy in the rest of the brain, but it's actually a stronger effect than age-related causes of brain atrophy itself. So put another way, when we talk about age-related brain shrinkage. That's not just due to the passage of time, but it's an artifact of having these other small vessel changes that themselves are inducing the loss of tissue. As we go to the next slide, we can also understand that our ability to track hemorrhages is limited on Hyperfine, as we see on the left compared to susceptibility weighted imaging in this same patient. And so I would not say that we would use Hyperfine to track ARIA-H. But certainly, the ability to track a more clinically important aspect of ARIA, I think makes the use of Hyperfine critically important for Alzheimer's imaging. However, as we go to the next slide, it's also important to understand that we have the ability to track not only FLAIR hyper-intensities and gliosis. But we can see atrophy noted on this right parietal gliosis in encephalomalacia on the Hyperfine T1. And so if we have a way of better evaluating the images and processing them using new artificial intelligence approaches we would have the ability to potentially track atrophy on a more sensitive level. And so as we go to the next slide, what you can see here is that there are tools also out of Harvard, thanks to the work of Dr. Eugenio Iglesias that can actually take Hyperfine scans and use artificial intelligence with a program called SynthSR to transform them into high-resolution scans, which would then allow for sensitive evaluation of regional atrophy. And in fact, at the upcoming Alzheimer's Association meeting, there will be multiple abstracts presented on Hyperfine applications in Alzheimer's, including using this program to better evaluate brain atrophy. And so we can use it also to evaluate small vessel disease more sensitively and these are all critical markers of brain health. And so moving to our next slide, and our final one right here, we can see that, in my opinion, Hyperfine is really important for scaling MRI for tracking Alzheimer's disease, especially ARIA-E with these new treatments. But it's also important for being able to evaluate other metrics of brain health, such as small vessel ischemic disease and brain atrophy and artificial intelligence tools like SynthSR will only make the ability to track these metrics more accurate over time. And for this reason, I think there's a lot of promise for applying Hyperfine in this space. Thank you so much for your time.

Thank you very much, Cyrus, as well as Taylor and Jamal, and we do have some questions from the participants in the audience. So why don't we go through some of those? The first question is directed to Jamal, and it's from Marie Thibault from BTIG. And basically, she's asking a couple of clinical practical questions. What was involved in the training of the X-ray techs any sort of workflows, changes that needed to be involved. And in addition, and this we could open up to anyone -- were there additional training requirements for the neuroradiologist interpreting the images from your standpoint?

Sure. So those are all very excellent questions. We did have our X-ray techs take a dedicated MR safety course. I was just looking up to see if I find the syllabus because they gave it to me at one point, and I couldn't find it. But basically, I think it was roughly 20 hours of instructions. It was a multi-day course that I think they took one day a week for several weeks. And basically, so they definitely got some specialized training in MR safety. But I wouldn't describe it as extensive, and we were able to train at least 5 or 6 techs now to do that. And it's much less than you would train an MR Tech in terms of years of schooling. So we felt that, that was adequate for them. The other question was workflow changes. So we did enable in Epic, there was a little bit of a workflow change in terms of the ordering clinician be able to put in a check box to say they wanted the study portable as opposed to high fields. So we've allowed the clinicians to make their decision and basically refer patients who they think is useful to the portable MRI. And we felt that, that was very useful and a low burden to introduce that. There also was training of the technologists by Hyperfine themselves, and Eddie, I'm sure can explain that. But basically, there were several days where they were on site and teaching the techs how to use the iPad that controls the scanner. It's very intuitive. It's very easy. It's much less burden than a standard MRI, where you basically can go in and tweak a lot of parameters like flip angles and SARs and things and TRs, repetition times. Basically, you just pull the sequences that you want over and you put them in a line and then press start. So it's pretty simple and they were able to teach all of our technologists that and the technologists could teach each other. So it's a very, very easy workflow for them to adopt basically.

Great. Thank you. There's another question. This is to Taylor. And it's from Ranjit from B. Riley. And the comment is, it's quite obvious to see some of the stroke signs and findings from MR. How often is it that they are not so obvious?

Yes. Thanks, Eddie, and thanks, Ranjit, for the question. Just to sort of emphasize, that is, in fact, the great advantage of MR-based detection of acute stroke as compared to non-con head CT. And -- and the idea is to leverage that strength of MR and the Swoop scanner to be able to detect it, we have been very interested in pushing the boundaries of what the Hyperfine scanner can detect, especially in terms of size. And we've gotten down to very, very small lesions at the size of about the pencil eraser volume, somewhere in that range. So as an acute stroke neurologist neurointensivists, having that degree of sensitivity and expecting that, that sensitivity will further improve as these additional sequence development and advancements that Hyperfine continues to push forward. I think it really highlights the value of the Swoop scanner in that setting. And just this one additional point, sort of the feasibility aspect of this is that to get the scan accomplished with these scans as a matter of 5 to 10 minutes to get a DWI. And so you really get an answer at the bedside and you're interacting in fact, with an iPad that the clinical provider at the bedside nurse has right there. And so you can make your decisions at that point of care and that really is transformational in terms of leveraging the technology for these types of acute decision-making for stroke.

Thanks Taylor. The next question comes from Young Li from Jefferies, and this is open to anyone who would like to answer. Basically, what percentage of the standard scans are your institutions you think can be potentially replaced or as additional imaging with the Swoop scanner and why?

All right. I'm going to try to take that and answer a couple of other questions. So in terms of the learning curve for our neuroradiologists, there's definitely a little bit of a learning curve. I think what I did was I put together a PowerPoint for our radiologists. And I think after they saw the pathology that helped them feel more comfortable all of our neuroradiologists are now comfortable. We do scan on the FLAIR in 2 directions, just to kind of add our clinical confidence. But -- so there is a bit of a learning curve, but it's definitely doable. The question of how many percent of our scans can be replaced, and it goes into one of the other questions that was asked. So in general, right now, this is FDA approved for noncontrast brain MRs. So when we're looking at candidates to scan, we look at anybody that has a noncontrast brain MRI ordered, and that's a candidate. Basically, the ones who are not, are the ones that have with contrast or the ones that have MRAs of their head and neck as well because those are things that we can't do with the Swoop. So just giving you a number, I would estimate maybe 50% -- little under 50% of our brain scans just because of the limitations of not using contrast and not MRAs.

Great. Thank you. Cyrus, anything you want to add to that? Or we could...

I want to say that if you're talking about standard scans. If by that, you mean high field MRI, you'll see in practice that we're going to get the high field MRI in conjunction with the portable MRI, but the portable MRI will be done earlier because of that increased access. I do think that there's a lot of potential for portable MRI to replace portable CT scans, which in my experience, I can really be really hard to interpret. And even when you do have a good quality portable CT scan, which is not always the case, it's much less sensitive for picking up really big things we care about, such as stroke. And so I could see in the future really this portable MRI approach of Hyperfine replacing a lot of portable CTs which, to give you some perspective, we do 80 million CTs in the United States a year, which is double the MRI statistics I cited earlier. So there's a big area in which you could have a lot more portable MRI contributions.

The next couple of questions relate to the CARE study as well as Alzheimer's and dementia. There is a question from Bradley Bowers from Mizuho. And basically, what Bradley is asking is, how significant do you view, Cyrus, that CARE PMR study in terms of the adoption of the technology and its impact upon scanning patients in the outpatient infusion centre environment?

I think the CARE PMR study is going to be the seminal study to address how exactly we have Hyperfine Swoop systems contribute to individuals undergoing treatment for Alzheimer's disease for a couple of reasons. One, it's very rigorously designed. And of course, in terms of the details of how and when it gets done, I would invite you to come to AAIC and talk to the study investigators directly because they'll have a better handle about that. But I do think the fact that they're tracking ARIA-E, which is I know it is more clinically important for understanding this, the fact that you have really strong collaboration across health systems that are experts of this from WashU to Harvard. You're going to have -- and then the focus on equity as well because if you have the ability -- and we haven't really talked about this, but if you have the ability to do Swoop MRI anywhere. You can give it access to communities that historically might not find it as easy to come to central medical centers for imaging. And so for those reasons, I think this is going to be the definitive study going forward.

Great. And then the last question is, again, from Ranjit. And basically, it gets to something at least in my mind, and I think the mind of a lot of neuroradiologist has the potential effect of being able to detect ARIA-E versus ARIA-H. I think you touched upon some of that and the clinical significant Cyrus. I wonder if you can kind of elaborate a little more?

Yes. I would say that I would have confidence in the ability of Hyperfine to detect ARIA-E, especially the peripheral edema. I think ARIA-H is going to be harder for detection just because the sequences that are needed to optimize the detection aren't currently available with the Hyperfine platform. But with ARIA-E, I would have confidence in using Hyperfine especially if I'm comparing it to a CT. Most people coming in for these sorts of acute symptomatic issues are going to get CTs first. I would much rather have a Hyperfine scan on board for that because it's not just going to tell me if there's edema there, but if there's stroke, if there's other bad stuff happening to the patient a lot more sensitively than a CT scan.

Great. Thanks. And actually, there's one more question that just snuck in, and that's from Simran Kaur who's from Wells Fargo. And the question is, how impactful are the software updates? So I know and I think Taylor kind of addressed this already with his 5 years of clinical experience, there's a significant impact, but I'll let Taylor answer that question.

Thank you, Eddie, Say and maybe Cyrus or Jamal can speak to the history of conventional MR better than I can. But at least in that context, when conventional MR first came around several decades ago, the rate of change was measured in units of decade. The rate of change at which Hyperfine has led the charge for ultra low-field MR is measured on the order of 6-month increments or annual increments and the degree of improvements, which I understand, largely relate to improved ML algorithms and imagery construction algorithms have had an absolutely dramatic impact. And I have every expectation that, that will continue. So it does make a large difference as a clinician at the bedside, when I have looked at the scans coming off the scanner and looking at the iPad, they really are now rivaling in quality and diagnostic quality as some contents for conventional market.

I'll just add. I was recently a panelist at the American Society of Neuro Radiology, where we had a panel on the Swoop. And one of the users presented some images that I think were from early on in the Swoop. And basically, they weren't comparable to what's currently available. So I think some of the resistance that may have been seen initially was due to initial image quality, which has really been improved at this point. So I think people can take another look because the images are definitely better. As Eddie pointed out in one of his talks are better than what people were initially using to diagnose using high-field MRIs. And they're definitely -- so the question right now is, can you answer the clinical question, not is this a better scan or equivalent scan to high field but rather can you answer the clinical question using the scanner. And I think in our study, 91% of the cases, we thought we could, so...

With that -- yes Cyrus.

I'll just add briefly -- sorry for the interruption that with artificial intelligence analysis and post-processing of Hyperfine Swoop images, the more of that you're going to see, the harder it's going to be to distinguish these images from high field MR images, which can not only be helpful for the adoption. So that's what I wanted to add.

Great. Thank you very much. So with that, I'd like to thank all of our participants for this fascinating and phenomenal webinar. I'd also like to add if there are any additional questions, feel free to address them to me. My e-mail address is on the website, but nonetheless, it's [email protected]. And with that, I'll turn it over to the Gilmartin Group for any final last words.

Thank you all for coming.

Okay. Short and sweet. Thank you, everyone. Take care.

The webinar is now over.